4 Defining a calculation scheme with the Python programming interface
5 ============================================================================
6 A YACS calculation scheme can be defined from a program written in the Python language (http://www.python.org/).
7 Refer to the `Python tutorial <http://docs.python.org/tut/tut.html>`_ for an introduction to the language.
9 The programming interface (API) is carried on three Python modules: pilot, SALOMERuntime and loader.
11 The SALOMERuntime module is used to initialise YACS for SALOME.
13 The loader module is used to create calculation schemes by loading files in the XML format.
15 The pilot module is used to create calculation schemes.
17 These modules must be imported at the beginning of the Python program and YACS must be initialised::
23 SALOMERuntime.RuntimeSALOME_setRuntime()
25 Before YACS modules can be imported, the environment must be correctly configured, as it will be if the
26 SALOME application is used. Otherwise, the PYTHONPATH environment variable has to be set to
27 <YACS_ROOT_DIR>/lib/pythonX.Y/site-packages/salome.
29 When you build your own Salome application and use your own modules and components (using YACSGEN for example), you may need to load
33 SALOMERuntime.RuntimeSALOME_setRuntime()
34 salome_runtime = SALOMERuntime.getSALOMERuntime()
37 mc = salome.naming_service.Resolve('/Kernel/ModulCatalog')
38 ior = salome.orb.object_to_string(mc)
39 session_catalog = salome_runtime.loadCatalog("session", ior)
40 salome_runtime.addCatalog(session_catalog)
45 Create a calculation scheme by loading an XML file
46 --------------------------------------------------------------
47 This is the easiest way of creating a calculation scheme. If there is a file conforming with the YACS syntax (see :ref:`schemaxml`),
48 then all that is necessary is to create an XML file loader and then to use its load method to obtain a calculation scheme object in Python.
50 The following shows the sufficient Python code to load an XML file::
52 xmlLoader = loader.YACSLoader()
54 p = xmlLoader.load("simple1.xml")
56 print "IO exception:",ex
59 Then, if the initialisation code and the loading code are put into a file named testLoader.py, proceed as follows::
63 to execute the program. The IOError exception can be raised by the loading operation principally if the file does not exist
64 or if it cannot be read. If no exception has been raised, it is necessary to make sure that the file analysis took place correctly.
65 This is done using the Logger object associated with the calculation scheme::
67 logger=p.getLogger("parser")
68 if not logger.isEmpty():
69 print "The imported file has errors :"
73 Finally, if the file analysis took place correctly, the validity of the scheme (completeness of connections, no unconnected
74 input port, etc.) has to be checked. This is done using the isValid method of the calculation scheme object, and
75 then the p.checkConsistency method of this object as below::
78 print "The schema is not valid and can not be executed"
79 print p.getErrorReport()
82 info=pilot.LinkInfo(pilot.LinkInfo.ALL_DONT_STOP)
83 p.checkConsistency(info)
84 if info.areWarningsOrErrors():
85 print "The schema is not consistent and can not be executed"
86 print info.getGlobalRepr()
90 If all these tests took place correctly, the scheme is ready to be executed (see :ref:`execpy`).
92 Create a calculation scheme from scratch
93 -------------------------------------------
94 We will use the same sequence as in :ref:`schemaxml`.
95 The first step is to obtain the runtime object that will be used for creation of objects making up the scheme, before they are created::
97 r = pilot.getRuntime()
99 Creating an empty scheme
100 ''''''''''''''''''''''''''''
101 An empty scheme is obtained using the createProc method of the runtime object with the name of the scheme as an argument::
105 The scheme object named “pr” was created. It is represented by the Python variable p.
107 Definition of data types
108 '''''''''''''''''''''''''''''''''
114 A basic type cannot be defined. These types are defined by YACS. However, it must be possible to retrieve a Python object
115 equivalent to a basic type so as to be able to subsequently create ports.
117 A basic data type is recovered using the getTypeCode method in the calculation scheme with the name of the type as an argument.
120 td=p.getTypeCode("double")
122 will obtain a double type (Python td object). Other basic types are obtained by::
124 ti=p.getTypeCode("int")
125 ts=p.getTypeCode("string")
126 tb=p.getTypeCode("bool")
127 tf=p.getTypeCode("file")
130 +++++++++++++++++++++
131 The createInterfaceTc method in the calculation scheme is used to define an object reference type.
132 This method accepts three arguments: the repository id of the corresponding SALOME object, the name of the type, and a
133 list of types that will be basic types of this type. If the repository id is equal to “”, the default value will be used.
135 The following is a minimal example for a reference definition of an object name Obj (default repository id, no basic type)::
137 tc1=p.createInterfaceTc("","Obj",[])
139 The same Obj type can be defined giving the repository id::
141 tc1=p.createInterfaceTc("IDL:GEOM/GEOM_Object","Obj",[])
143 A list of basic types is also provided so as to define a reference object type derived from another type.
145 The following gives a definition of the MyObj type derived from the Obj type::
147 tc2=p.createInterfaceTc("","MyObj",[tc1])
150 +++++++++++++++++++++
151 The createSequenceTc method in the calculation scheme is used to define a sequence type.
152 This method accepts three arguments, namely the repository id, the type name, and the type of elements in the sequence.
153 There is generally no point in specifying the repository id. The value “” will be given.
155 The following gives an example definition of the seqdbl double sequence type::
157 tc3=p.createSequenceTc("","seqdbl",td)
159 td is the double type that is obtained as above in the section on :ref:`basictypes`.
161 A sequence type of sequence is defined as follows::
163 tc4=p.createSequenceTc("","seqseqdbl",tc3)
165 A reference sequence type is defined as follows::
167 tc5=p.createSequenceTc("","seqobj",tc1)
171 A structure type is defined using the createStructTc method in the calculation scheme.
172 This method accepts two arguments, namely the repository id and the type name. For standard use, the repository id is
173 equal to the value “”. The structure type is the only type that is defined in two steps. It is created empty after
174 calling the createStructTc method. Its members are then defined by adding them with the addMember method.
176 The following shows an example definition of an s1 type structure with 2 members (m1 and m2) of the double and double sequence types::
178 ts1=p.createStructTc("","s1")
179 ts1.addMember("m1",td)
180 ts1.addMember("m2",tc3)
182 Retrieve predefined types
183 +++++++++++++++++++++++++++++++++
184 By default, YACS only defines the basic types. If more predefined types are required, they must be requested from SALOME.
185 These other predefined types are contained in module catalogs such as GEOM or SMESH.
187 The following code sequence is used to obtain an image of SALOME catalogs in YACS::
190 cata=r.loadCatalog("session",
191 "corbaname::localhost:2810/NameService#Kernel.dir/ModulCatalog.object")
192 except CORBA.TRANSIENT,ex:
193 print "Unable to contact server:",ex
194 except CORBA.SystemException,ex:
195 print ex,CORBA.id(ex)
197 The SALOME application must be running before the catalog is accessible.
198 Predefined types are then accessible in the cata._typeMap dictionary.
199 If the name of the required type is known (for example ‘GEOM_Shape’), it is obtained as follows::
201 tgeom=cata._typeMap['GEOM_Shape']
205 Add a type into the scheme types dictionary
206 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
207 Some operations require that types are defined in the scheme dictionary. Proceed as follows if you want to add a type
208 into the dictionary, for example for the seqobj type defined above::
210 p.typeMap["seqobj"]=tc5
212 where the type name is the dictionary key and the type is the value.
214 Definition of elementary calculation nodes
215 ''''''''''''''''''''''''''''''''''''''''''''''
220 +++++++++++++++++++++
221 Several steps are used to define a script node in a given context (for example the calculation scheme).
222 The first step consists of creating the node object by calling the runtime createScriptNode method.
223 This method uses 2 arguments, the first of which in standard use must be equal to “” and the second is the node name.
224 The following is an example to create node node1::
226 n=r.createScriptNode("","node1")
228 The second step consists of attaching the node to its definition context by calling the edAddChild method for the context object.
229 This method has one argument, namely the node to be attached. The following is an example of the attachment of the node node1
230 to the calculation scheme::
234 Warning: the name of the method to be used depends on the type of context node. We will see which method should be used for other
237 The third step consists of defining the Python script associated with the node. This is done using the setScript method for the node
238 with a character string argument that contains the Python code. The following shows an example definition of the associated code::
240 n.setScript("p1=p1+2.5")
242 The fourth step consists of defining input and output data ports. An input port is created by calling the edAddInputPort method
243 for the node. An output port is created by calling the edAddOutputPort method for the node.
244 These two methods have two arguments: the port name and the port data type. The following is an example creating a double
245 type input port p1 and a double type output port p1::
247 n.edAddInputPort("p1",td)
248 n.edAddOutputPort("p1",td)
250 Our node is now fully defined with its name, script, ports and context. It retrieves the double in the input port p1, adds 2.5 to it
251 and puts the result into the output port p1.
253 If you want to execute your script node on a remote container, you have to set the execution mode of the node to **remote**
254 and to assign a container (see :ref:`py_container` to define a container) to the node as in the following example::
256 n.setExecutionMode("remote")
257 n.setContainer(cont1)
259 The default option for the execution mode is **local** where the node will run
260 on the same container as the scheme executor.
265 ++++++++++++++++++++++
266 The same procedure is used to define a function node. The only differences apply to creation, in using the createFuncNode
267 method and defining the function: the setFname method must also be called to give the name of the function to be executed.
268 The following is a complete example for the definition of a function node that is functionally identical to the previous script node::
270 n2=r.createFuncNode("","node2")
278 n2.edAddInputPort("p1",td)
279 n2.edAddOutputPort("p1",td)
281 If you want to execute your function node on a remote container, you have to set the execution mode of the node to **remote**
282 and to assign a container (see :ref:`py_container` to define a container) to the node as in the following example::
284 n2.setExecutionMode("remote")
285 n2.setContainer(cont1)
287 The default option for the execution mode is **local** where the node will run
288 on the same container as the scheme executor.
293 ++++++++++++++++++++++++++
294 There are two definition forms for a SALOME service node.
296 The first form in which the component name is given, uses the createCompoNode method to create the node. The name of the
297 component is given as an argument to the setRef method for the node. The service name is given as an argument for the
298 setMethod method for the node. The remainder of the definition is exactly the same as for the previous Python nodes.
300 The following is an example of a node that calls the makeBanner service for a PYHELLO component::
302 n3=r.createCompoNode("","node3")
305 n3.setMethod("makeBanner")
306 n3.edAddInputPort("p1",ts)
307 n3.edAddOutputPort("p1",ts)
309 The second form is used to define a node that uses the same component as another node uses the createNode method of this other node.
310 This method only has one argument, which is the node name.
311 The remainder of the definition is identical to the definition for the previous form.
313 The following gives an example of a service node that makes a second call to the makeBanner service for the same component
314 instance as the previous node::
316 n4=n3.createNode("node4")
318 n4.setMethod("makeBanner")
319 n4.edAddInputPort("p1",ts)
320 n4.edAddOutputPort("p1",ts)
322 Definition of connections
323 ''''''''''''''''''''''''''''
324 Obtaining a node port
325 ++++++++++++++++++++++++++++
326 Before links can be defined, it is almost always necessary to have Python objects representing the output port to be
327 connected to the input port. There are two ways of obtaining this object.
329 The first way is to retrieve the port when it is created using the edAddInputPort and edAddOutputPort methods.
330 For example, we can then write::
332 pin=n4.edAddInputPort("p1",ts)
333 pout=n4.edAddOutputPort("p1",ts)
335 pin and pout are then the objects necessary to define links.
337 The second way is to interrogate the node and ask it for one of its ports by its name.
338 This is done using the getInputPort and getOutputPort methods.
339 pin and pout can then be obtained as follows::
341 pin=n4.getInputPort("p1")
342 pout=n4.getOutputPort("p1")
345 ++++++++++++++++++++++++++++
346 The edAddCFLink method for the context is used to define a control link between two nodes, transferring the two nodes to be
347 connected to it as arguments. For example, a control link between nodes n3 and n4 will be written::
351 Node n3 will be executed before node n4.
354 ++++++++++++++++++++++++++++
355 The first step in defining a dataflow link is to obtain port objects using one of the methods described above.
356 Then, the edAddLink method links an output port to an input port::
358 pout=n3.getOutputPort("p1")
359 pin=n4.getInputPort("p1")
360 p.edAddLink(pout,pin)
362 Most of the time, when you need a dataflow link between two ports, you also need a control link between the nodes
363 of the ports. In this case you can use the method edAddDFLink::
365 pout=n3.getOutputPort("p1")
366 pin=n4.getInputPort("p1")
367 p.edAddDFLink(pout,pin)
369 edAddDFLink is equivalent to edAddCFLink followed by edAddLink.
371 Initialising an input data port
372 '''''''''''''''''''''''''''''''''''''''''''''''
373 An input data port is initialised firstly by obtaining the corresponding port object. There are then two methods of initialising it.
375 The first method initialises the port with a value encoded in XML-RPC. The edInitXML method for the port is then used.
376 The following is an example that initialises the port with the integer value 5::
378 pin.edInitXML("<value><int>5</int></value>")
380 The second method initialises the port with a Python value. The edInitPy method is then used.
381 The following is an example that initialises this port with the same value::
385 Specific methods can also be used for basic types:
387 - ``edInitInt`` for the int type
388 - ``edInitDbl`` for the double type
389 - ``edInitBool`` for the bool type
390 - ``edInitString`` for the string type
392 First example starting from the previous elements
393 '''''''''''''''''''''''''''''''''''''''''''''''''''
394 By collecting all previous definition elements, a complete calculation scheme identical to that given in the :ref:`schemaxml` chapter
395 will appear as follows::
401 SALOMERuntime.RuntimeSALOME_setRuntime()
402 r = pilot.getRuntime()
404 ti=p.getTypeCode("int")
405 td=p.getTypeCode("double")
407 n1=r.createScriptNode("","node1")
409 n1.setScript("p1=p1+10")
410 n1.edAddInputPort("p1",ti)
411 n1.edAddOutputPort("p1",ti)
413 n2=r.createScriptNode("","node2")
415 n2.setScript("p1=2*p1")
416 n2.edAddInputPort("p1",ti)
417 n2.edAddOutputPort("p1",ti)
419 n4=r.createCompoNode("","node4")
422 n4.setMethod("echoDouble")
423 n4.edAddInputPort("p1",td)
424 n4.edAddOutputPort("p1",td)
429 p.edAddLink(n1.getOutputPort("p1"),n2.getInputPort("p1"))
430 p.edAddLink(n1.getOutputPort("p1"),n4.getInputPort("p1"))
431 #initialisation ports
432 n1.getInputPort("p1").edInitPy(5)
434 Definition of composite nodes
435 '''''''''''''''''''''''''''''''''
441 A block is defined using the runtime createBloc method transferring the Block name to it as an argument. The node is then
442 attached to its definition context as an elementary node. The following is an example Block definition in a calculation scheme::
447 Once the block has been created, all nodes and links possible in its context can be added.
448 Repeating a part of the example above, we will get::
450 n1=r.createScriptNode("","node1")
452 n1.setScript("p1=p1+10")
453 n1.edAddInputPort("p1",ti)
454 n1.edAddOutputPort("p1",ti)
455 n2=r.createScriptNode("","node2")
457 n2.setScript("p1=2*p1")
458 n2.edAddInputPort("p1",ti)
459 n2.edAddOutputPort("p1",ti)
460 b.edAddDFLink(n1.getOutputPort("p1"),n2.getInputPort("p1"))
466 A Forloop is defined using the runtime createForLoop method, transferring the node name to it as an argument.
467 The node is then attached to its definition context. The following is an example ForLoop definition in a calculation scheme::
469 l=r.createForLoop("l1")
472 The number of iterations in the loop to be executed will be initialised using the “nsteps” port that is initialised
473 with an integer. For example::
475 ip=l.getInputPort("nsteps")
478 There is a special method for obtaining the “nsteps” port for the loop, namely edGetNbOfTimesInputPort. Therefore, it can also be
481 ip=l.edGetNbOfTimesInputPort()
484 Finally, a method called edSetNode will be used in the context of a loop, instead of the edAddChild method, so as to add one (and only one) node.
485 The following is a small example definition of a node inside a loop::
487 n1=r.createScriptNode("","node1")
489 n1.setScript("p1=p1+10")
490 n1.edAddInputPort("p1",ti)
491 n1.edAddOutputPort("p1",ti)
497 WhileLoop node is defined in practically the same way as a ForLoop node. The only differences apply to creation and assignment
498 of the end of loop condition. The createWhileLoop method is used for creation. The “condition” port is used for the condition.
499 If looping takes place on a node, it is important to use a data link instead of a dataflow link.
500 The following is an example of WhileLoop node definition with a Python script internal node.
501 The condition is initialised to True and is then changed to False by the internal node. This results in a link loop::
503 wh=r.createWhileLoop("w1")
505 n=r.createScriptNode("","node3")
507 n.edAddOutputPort("p1",ti)
509 cport=wh.getInputPort("condition")
510 cport.edInitBool(True)
511 p.edAddLink(n.getOutputPort("p1"),cport)
513 There is a special method for obtaining the loop “condition” port: edGetConditionPort.
519 A ForEach node is basically defined in the same way as any other loop node. There are several differences.
520 The node is created with the createForEachLoop method that has an additional argument, namely the data type managed by the ForEach.
521 The number of ForEach branches is specified with the “nbBranches” port. The collection on which the ForEach iterates is managed by
522 connection of the “evalSamples” and “SmplsCollection” ports.
524 The following is an example definition of the ForEach node with a Python script internal node that increments
525 the element of the collection by 3::
527 fe=r.createForEachLoop("fe1",td)
529 n=r.createScriptNode("","node3")
530 n.setScript("p1=p1+3.")
531 n.edAddInputPort("p1",td)
532 n.edAddOutputPort("p1",td)
534 p.edAddLink(fe.getOutputPort("evalSamples"),n.getInputPort("p1"))
535 fe.getInputPort("nbBranches").edInitPy(3)
536 fe.getInputPort("SmplsCollection").edInitPy([2.,3.,4.])
538 Special ports for the ForEach can be obtained using the following methods instead of getInputPort and getOutputPort:
540 - edGetNbOfBranchesPort for the “nbBranches” port
541 - edGetSamplePort for the “evalSamples” port
542 - edGetSeqOfSamplesPort for the “SmplsCollection” port
548 A switch node is defined in several steps. The first two steps are creation and attachment to the context node.
549 The node is created by calling the runtime createSwitch method with the name of the node as an argument. The node is attached
550 to the context node by calling the edAddChild method for a scheme or a block or edSetNode for a loop node.
552 The following is an example of a creation followed by an attachment::
554 sw=r.createSwitch("sw1")
557 The next step is to create an internal elementary or composite node by case. The node for the default case is attached to
558 the switch using the edSetDefaultNode method. Nodes for other cases are attached to the switch using the edSetNode method, in
559 which the first argument is equal to the value of the case (integer) and the second argument is equal to the internal node.
561 The following is an example of a switch with one script node for case “1” and another script node for the “default” case
562 and a script node to initialise an exchanged variable::
565 n=r.createScriptNode("","node3")
566 n.setScript("p1=3.5")
567 n.edAddOutputPort("p1",td)
570 sw=r.createSwitch("sw1")
572 nk1=r.createScriptNode("","ncas1")
573 nk1.setScript("p1=p1+3.")
574 nk1.edAddInputPort("p1",td)
575 nk1.edAddOutputPort("p1",td)
577 ndef=r.createScriptNode("","ndefault")
578 ndef.setScript("p1=p1+5.")
579 ndef.edAddInputPort("p1",td)
580 ndef.edAddOutputPort("p1",td)
581 sw.edSetDefaultNode(ndef)
582 #initialise the select port
583 sw.getInputPort("select").edInitPy(1)
584 #connection of internal nodes
585 p.edAddDFLink(n.getOutputPort("p1"),nk1.getInputPort("p1"))
586 p.edAddDFLink(n.getOutputPort("p1"),ndef.getInputPort("p1"))
588 The edGetConditionPort method can be used instead of getInputPort, to obtain the special “select” port for the Switch.
590 .. _py_optimizerloop:
595 The following is an example of OptimizerLoop with one python script as internal node. The algorithm
596 is defined by the class async in the python module myalgo2.py::
598 ol=r.createOptimizerLoop("ol1","myalgo2.py","async",True)
600 n=r.createScriptNode("","node3")
602 n.edAddInputPort("p1",td)
603 n.edAddOutputPort("p1",ti)
605 ol.getInputPort("nbBranches").edInitPy(3)
606 ol.getInputPort("algoInit").edInitPy("coucou")
607 p.edAddLink(ol.getOutputPort("evalSamples"),n.getInputPort("p1"))
608 p.edAddLink(n.getOutputPort("p1"),ol.getInputPort("evalResults"))
612 Definition of containers
613 ''''''''''''''''''''''''''''
615 This example shows how to add a container to a scheme::
617 c1=p.createContainer("MyContainer")
619 A property is added to a container using its setProperty method that uses 2 arguments (character strings).
620 The first is the property name. The second is its value.
621 The following is an example of how to set constraints on the container::
623 c1.setProperty("container_name","FactoryServer")
624 c1.setProperty("hostname","localhost")
625 c1.setProperty("mem_mb","1000")
627 Once the containers have been defined, SALOME components can be placed on this container. The first step to place the component
628 of a SALOME service node is to obtain the component instance of this service node using the getComponent method for this node.
629 The previously defined container is then assigned to this component instance using the setContainer method of the component instance.
631 If it is required to place the SALOME service defined above (node “node3”) on
632 “MyContainer”, we will write::
634 n3.getComponent().setContainer(c1)
636 It is also possible to place python nodes on containers, but the code is a
637 little different (see :ref:`pyscript`)::
639 n1.setExecutionMode("remote")
642 Since SALOME v7.5, there is a new type of container:
643 *Homogeneous Pool of SALOME containers* (HP container).
644 It is possible to create this type of container this way::
646 my_hp_cont=r.createContainer("MyHPCont","HPSalome")
648 - "MyHPCont" : name of the container. Same result as my_hp_cont.setName("MyHPCont").
649 - "HPSalome" : type of container. Possible values are "HPSalome" (for a HP container)
650 or "Salome" (for a classic container).
653 '''''''''''''''''''''''''''
654 A property is added to an elementary or composite node (or is modified) using its setProperty method that has two
655 arguments (character strings). The first is the name of the property. The second is its value.
656 The following is an example for the previous node “node3”::
658 n3.setProperty("VERBOSE","2")
660 Datastream connections
661 '''''''''''''''''''''''''''
662 Datastream connections are only possible for SALOME service nodes as we have seen in :ref:`principes`.
663 We firstly need to define the datastream ports in the service node. An input datastream port is defined using
664 the edAddInputDataStreamPort method. An output datastream port is defined using the edAddOutputDataStreamPort method.
665 These methods use the port name and the datastream type as arguments.
667 Some datastream ports (for example CALCIUM ports) must be configured with properties. The port setProperty method will
668 be used to configure them.
669 The following is an example definition of the SALOME service node with datastream ports. This is the DSCCODC component
670 located in the DSCCODES module in the EXAMPLES base. The datastream ports are of the “CALCIUM_integer” type
671 with time dependency::
673 calcium_int=cata._typeMap['CALCIUM_integer']
674 n5=r.createCompoNode("","node5")
678 pin=n5.edAddInputDataStreamPort("ETP_EN",calcium_int)
679 pin.setProperty("DependencyType","TIME_DEPENDENCY")
680 pout=n5.edAddOutputDataStreamPort("STP_EN",calcium_int)
681 pout.setProperty("DependencyType","TIME_DEPENDENCY")
683 Once the service nodes have been provided with datastream ports, all that remains is to connect them.
684 This connection is made using the edAddLink method for the context node in the same way as for data links.
685 The only difference is the type of ports transferred as arguments.
687 To complete our example, we will define a second service node and connect the datastream ports for these services::
689 n6=r.createCompoNode("","node6")
693 pin=n6.edAddInputDataStreamPort("ETP_EN",calcium_int)
694 pin.setProperty("DependencyType","TIME_DEPENDENCY")
695 pout=n6.edAddOutputDataStreamPort("STP_EN",calcium_int)
696 pout.setProperty("DependencyType","TIME_DEPENDENCY")
697 p.edAddLink(n5.getOutputDataStreamPort("STP_EN"),n6.getInputDataStreamPort("ETP_EN"))
698 p.edAddLink(n6.getOutputDataStreamPort("STP_EN"),n5.getInputDataStreamPort("ETP_EN"))
700 Other elementary nodes
701 '''''''''''''''''''''''''''''''
704 A SalomePython node is defined in practically exactly the same way as a :ref:`pyfunc`. The runtime createSInlineNode method is used
705 instead of the createFuncNode and information about placement on a container is added in the same way as for a
706 SALOME service node (setContainer method).
708 The following is an example similar to that given in :ref:`schemaxml`::
710 n2=r.createSInlineNode("","node2")
717 print __container__from__YACS__
718 machine,container=__container__from__YACS__.split('/')
719 param={'hostname':machine,'container_name':container}
720 compo=salome.lcc.LoadComponent(param, "PYHELLO")
721 print compo.makeBanner(p1)
725 n2.edAddInputPort("p1",ts)
726 n2.getComponent().setContainer(c1)
732 A DataIn node is defined using the runtime createInDataNode method. It uses two arguments, the first of which must be “” and
733 the second the node name. Node data are defined by adding output data ports to it using the edAddOutputPort method
734 and transferring the data name and its type to it as arguments.
735 The value of the data is initialised using the port setData method thus created by transferring the value encoded in
736 XML-RPC to it (see :ref:`initialisation`).
738 The following is an example of the DataIn node that defines 2 double type data (b and c) and one file type data (f)::
740 n=r.createInDataNode("","data1")
742 pout=n.edAddOutputPort('a',td)
743 pout.setData("<value><double>-1.</double></value>")
744 pout=n.edAddOutputPort('b',td)
745 pout.setData("<value><double>5.</double></value>")
746 pout=n.edAddOutputPort('f',tf)
747 pout.setData("<value><objref>f.data</objref></value>")
749 A value can be directly assigned to a data with a Python object, using the setDataPy method. Example for a sequence::
751 pout.setDataPy([1.,5.])
757 A DataOut node is defined using the runtime createOutDataNode method. It uses two arguments, the first of which
758 must be “” and the second the node name . Node results are defined by adding input data ports to it using the edAddInputPort
759 method with the result name and its type as arguments. The results are saved in a file using the node setRef method with the
760 file name as an argument.
761 A result file is copied into a local file using the setData method for the port corresponding to the result with the
762 file name as an argument.
764 The following is an example of the DataOut node that defines different types (double, int, string, doubles vector, file) of
765 results (a, b, c, d, f) and writes the corresponding values in the g.data file.
766 The result file will be copied into the local file myfile::
768 n=r.createOutDataNode("","data2")
771 n.edAddInputPort('a',td)
772 n.edAddInputPort('b',ti)
773 n.edAddInputPort('c',ts)
774 n.edAddInputPort('d',tc3)
775 pin=n.edAddInputPort('f',tf)
776 pin.setData("monfich")
782 A StudyIn node is defined using the runtime createInDataNode method. It uses two arguments, the first of which must be “study”
783 and the second the node name. Node data are defined by adding output data ports using the edAddOutputPOrt method, transferring
784 the name of the data and its type as arguments. The data is initialised with the reference in the study, using the setData method
785 for the port thus created, transferring a character string to it containing either the SALOME Entry or the path in the study
788 The following is an example of the StudyIn node that defines 2 GEOM_Object type data (a and b). The study is assumed to be
789 loaded into memory by SALOME. Data a is referenced by one SALOME Entry. Data b is referenced by a path in the
790 study tree structure::
792 n=r.createInDataNode("study","study1")
794 pout=n.edAddOutputPort('a',tgeom)
795 pout.setData("0:1:1:1")
796 pout=n.edAddOutputPort('b',tgeom)
797 pout.setData("/Geometry/Sphere_1")
803 A StudyOut node is defined using the runtime createOutDataNode method. It uses two arguments, the first of
804 which must be “study” and the second the node name. The name of the file in which the study will be
805 saved is specified using the node SetRef method with the file name as an argument.
806 The node results are defined by adding input data ports to it using the edAddInputPort method, transferring the data name
807 and type as arguments. The setData method for the port is used to associate the entry into the study to the result, transferring
808 a character string to it that contains either the SALOME Entry or the path in the study tree structure.
810 The following contains an example of the StudyOut node that defines two GEOM_Object type results (a and b).
811 Result a is referenced by a SALOME Entry. The result b is referenced by a path.
812 The complete study is saved in the study1.hdf file at the end of the calculation::
814 n=r.createOutDataNode("study","study2")
815 n.setRef("study1.hdf")
817 pout=n.edAddInputPort('a',tgeom)
818 pout.setData("0:1:2:1")
819 pout=n.edAddInputPort('b',tgeom)
820 pout.setData("/Save/Sphere_2")
822 Save a calculation scheme in an XML file
823 ------------------------------------------------------
824 A calculation scheme is saved in a file in the XML format using the saveSchema method for the calculation
825 scheme, transferring the file name to it as an argument. Before a calculation scheme constructed under Python
826 can be saved in a consistent form in an XML file, all types defined in Python have to be added to the scheme types
827 dictionary (see :ref:`typedict`). The save will not do this automatically.
829 Proceed as follows to save the scheme p constructed above in the myscheme.xml file::
831 p.saveSchema("monschema.xml")
833 The file thus obtained can then be loaded as in :ref:`loadxml`.
835 Several useful operations
836 ------------------------------
838 Finding a node by its name
839 '''''''''''''''''''''''''''''''''''
840 A node (Python object) can be found, when all that is available is the calculation scheme object and
841 the absolute name of the node, by calling the scheme getChildByName method, transferring the absolute name to it.
843 To find the Python script node defined in :ref:`pyscript`::
845 n=p.getChildByName("node1")
847 To find node “node1” node in block “b1”::
849 n=p.getChildByName("b1.node1")
851 This operation can also be used starting from a composite node provided that the relative node name is used.
852 The previous example can be rewritten::
854 n=b.getChildByName("node1")
856 Finding a port by its name
857 '''''''''''''''''''''''''''''''''''
858 The first step to find a node port by its name is to retrieve the node by its name. An input data port is then found
859 using the getInputPort method, and an output data port is found using the getOutputPort method.
861 The following is an example starting from the previous node n::
863 pin=n.getOutputPort("p1")
864 pout=n.getInputPort("p2")
866 Obtaining a port value
867 '''''''''''''''''''''''''''''''''''
868 The value of a port is obtained using its getPyObj method. For example::
871 print pout.getPyObj()
873 Obtaining the state of a node
874 '''''''''''''''''''''''''''''''''''
875 The state of a node is obtained using its getEffectiveState method (see possible values in :ref:`etats`).
877 Removing a node from its context
878 '''''''''''''''''''''''''''''''''''
879 A node can be removed from its context node using a context method. The method name will be different
880 depending on the context type.
882 - For a block or a calculation scheme, the edRemoveChild method will be used with the node to be removed as an argument::
886 - For a loop (ForLoop, WhileLoop or ForEachLoop) the edRemoveNode method will be used without any argument::
890 - The edRemoveChild method will be used for a Switch, with the internal node concerned as an argument::
892 sw.edRemoveChild(nk1)